Come on in and sit back and relax. You’re listening to Episode 201 of the WealthTech Today podcast. I’m your host, Craig Iskowitz, founder of Ezra Group Consulting. This podcast features interviews, news and analysis on the trends and best practices, all about wealth management technology.
My guest for this episode is Mike Sha, CEO and founder of SigFig. You may have heard Mike Sha on our program before, we’ve also written a couple of blog posts where he has been on panels at conferences, where we were together speaking at conferences, so we go back a bit. Mike launched SigFig in 2011, it’s been 13 years. Previously, he was the CEO and co founder of a website called Wiki Invest, which then they rolled into SigFig.
SigFig was one of the early providers of direct to consumer wealth management tools and technologies was very, very popular. They got a deal with Yahoo back in the day, and then they added b2b and now they’re one of the biggest b2b digital advice vendors. Before SigFig and Wiki Invest. Mike worked at Amazon, from 2001 I think Amazon was founded in ’99, they went public in 99? He was one of the earlier employees 2001, 2005 Mike and I always hit it off because we’re both comp-sci people. He has a master’s though I don’t have a master’s, just a lowly undergrad Comp Sci majors am I.
SigFig has raised a total of $120 million in funding, and their Series E round was in June of 2018. Earlier this year back in February of 2023 SigFig announced the deal with Santander Bank, where they will be using SigFig software to build their digital advice channel.
But before we get started, if you are an executive at a broker-dealer, enterprise RIA, family office or a TAMPs, your tech debt is holding you back. Your old software platforms are rusty and falling apart and they need either a complete overhaul or to be replaced entirely. Your disparate systems don’t communicate with each other and it’s driving your operations staff and advisors crazy with manual processes and other errors. If this describes your company, you should run, not walk to our website, EzraGroup.com and fill out the Contact Us form on the home page. Our experienced team can evaluate your technology ecosystem, deliver targeted recommendations, optimize your existing systems and operations, or run an RFP or RFI to help you implement new software to help take your firm to the next level.
- The Evolution of SigFig
- How Does AI Fit Into SigFig?
- AI Trends in Wealth Management
- New Product: Engage
- Technology and Client Engagement
- Early Adopters vs Pragmatists
Craig: I am excited to introduce our next guest. It is Mike Sha, founder and CEO of SigFig. Mike, thanks for coming on the program.
Mike: My pleasure. Thanks for having me.
Craig: It’s exciting that you’re that you’re here. I’m glad we could work it out. I know you’re very busy. Where are you calling in from?
Mike: San Francisco. Me and four other people all in the area.
Craig: There’s some people who stayed, I’m pretty sure there’s a couple people still hanging around, still dealing with all the issues. We were talking earlier you guys have gone fully remote.
The Evolution of SigFig
Mike: We have. We have a global team split across quite a few countries. We were already very remote friendly before the pandemic but of course, the pandemic opened up lots of new talent pools and we love that. It’s been great to be able to hire the best people in the world from wherever they want to be.
Craig: I would agree. Ezra Group has always been remote, so we felt the same way. COVID didn’t bother us. We were just talking about how I was at a happy hour at your San Francisco office, that was a beautiful office. I’m sorry, you closed that.
Mike: You miss seeing people every day. But we’ve tried to augment being remote with obviously fun, interactive. Activities virtually, but even more importantly, being able to have team members get together in person. Being able to build relationships obviously is still an important part of having a good time and doing good work.
Craig: I would agree with you 100%. Let’s dive in. Can you give the audience a 30-second overview elevator pitch of SigFig?
Mike: Sure, we work with quite a few financial institutions, basically helping to accelerate their digital transformation and their whole move to digital and adoption of technology. Historically, many firms know of us, because we were one of the first b2b robo advisors. Back when robo advice was kind of a big thing. Over the years, our product line has expanded quite a bit. So we built a lot of technology for actually traditional financial advisors to use, and that’s a big part of who we are and what we do.
Mike: We’ve also built a bunch of new product lines in the kind of virtual collaboration and sales effectiveness space and trying to drive lead gen and growth. We’ve tapped into lots of new areas, but big part of who we are is we believe that you know the the usage of technology in the wealth management industry is nowhere near where it could be. And the best way to have a big impact on the industry is to actually work with traditional financial institutions, put better products, better technologies in their hands so that all of the clients that they serve can access the latest and greatest happening in the tech space.
Mike: Everyone wants the latest and greatest, Mike. I haven’t met a client of mine doesn’t want the latest and greatest. The issue is, how do they deploy it? How do they integrate it into their existing technology and get their advisors to use it? Now adoption is always an issue.
How Does AI Fit Into SigFig?
Craig: One thing we’ve been talking about a lot is and everyone’s been talking about is artificial intelligence. So I want to touch on that a little bit, can you talk about your thoughts on AI what is what is SigFig doing, are you just kicking the tires, how do you see AI fitting into your product line?
Mike: It’s a great question. I think a lot of people’s exposure to AI comes from what they see in the in the media. The first thing to talk about is you know, when you look at AI, what does it mean? I think one of the common misperceptions is that because ChatGPT and conversational interfaces, and large language models, that’s what’s all the rage, that’s what’s getting all the media attention. A lot of people equate those things to AI when it turns out that is a much bigger umbrella of technologies.
Mike: You think about natural language processing being an area you think about the whole kind of decision support space, neural networks. There are just so many subfields in AI. Even though the conversational side of the world has gotten a lot of the world’s attention right now, I actually think some of the most interesting applications in wealth management go well beyond you know, kind of what ChatGPT does and chat bots and things like that. The first thing to think to kind of recognize, at least from our perspective, is that the deployment and development of AI technologies and wealth will leverage a much broader universe of underlying technologies than just some of the language stuff that’s getting a lot of attention.
Mike: Another kind of I’d say area to kind of think about is, what are the most important problems to solve with AI? When I look at the way the world has been talking about AI, I think it falls into a few buckets that actually have some pretty direct translations to the wealth management industry. You know, I think the big three themes that we think about are one, instead of thinking about AI replacing jobs, actually, how does AI help augment the capabilities of humans and providers? So that’s one big theme.
Mike: The second theme is around automation. There are going to be plenty of areas where there’s an opportunity to leverage AI to create a lot more efficiency and a lot more scalability through automation. And then the third angle is about how AI can help augment the client experience. Right? We’re all here to serve clients and there’s no doubt that AI will be used to improve the actual client experience. In each of those buckets, very different applications, very different use cases. But I think we’re big believers that AI is absolutely a super important trend happening in technology and will have very direct benefits to the wealth management industry.
AI Trends Across Wealth Management
Craig: You mentioned a much broader universe of technology. More than just ChatGPT, can you give us some examples of what other technology you see being incorporated into AI solutions?
Mike: So let’s start with some of the more important buckets. There are some fields in AI they’re called, depending on which subfield sometimes people refer to as expert systems. Some people refer to it as machine learning. Some people refer to it as neural networks. All of these areas have some amount of similarity, which is essentially using artificial intelligence and technology to improve decision making systems. And so if you think about the importance of making good decisions in the investment space, giving good advice, the right recommendations, optimal financial planning, even we’re actually in the early days of robo advice, right? People thought of robo advice as algorithmic asset management.
Mike: It shocked me. I was not an industry insider. So the earlier part of my career I spent at Amazon, not inside of a bank. And as I got into the industry, it’s just amazing how much middle and back office work is still done manually. In all of those areas, AI will be incredibly impactful in both increasing quality and scalability and consistency with how business is done, while at the same time, likely reducing cost through automation and so that’s kind of a big general area.
Mike: Now, I think when you get down to it like well, what are the applications, is it about portfolio rebalancing? Is it about principle review? Is it about algorithms that helped give financial advice like these are all going to be areas that are going to be relevant applications of AI that go beyond kind of conversational LLMs. You’ve got other kind of buckets of things. Automated data extraction is kind of an interesting area. There’s a lot of paperwork involved in the investment business and it turns out that AI is going to be used quite effectively to actually ingest and analyze and optimize paperwork. When you talk to advisors, one of things that you hear is that paperwork is just so painful. So if that’s a day in and day out, kind of part of doing business, how can AI be used to actually make all of that kind of paperwork and workflow, a lot more easy for all involved, whether you’re the advisor or the client.
New Product: Engage
Craig: Is SigFig working on any specific areas of automated data extraction is number of startups that are automating data extraction, whether estate planning forms, tax forms, and others. So are there any particular areas you guys looking at?
Mike: One of our product lines is called Engage and with Engage, I’ll give you the setup of it, Engage for us was a product that was born during the pandemic, and started out as a product designed to help facilitate virtual collaboration. So imagine you’re a financial advisor and you’re meeting with a client. If you are doing any virtual collaboration today, it’s probably using like a video conferencing tool, Zoom or Teams or something like that. We ended up building Engage because we believe that the power of virtual collaboration in finance, actually extends well beyond video conferencing. And so we kind of built this tool that integrated technology to actually facilitate collaboration with clients.
Mike: So let’s say you’re doing like a financial plan or you’re doing needs discovery. You know, these are kind of interactions that advisors have with clients traditionally, through verbal conversation, you kind of talked to clients. But it turns out, you can leverage technology to make that conversation a lot more interactive, a lot more collaborative. That was part of what we did.
Mike: Another part of what we did was built around workflow so how can you make day to day common workflows that advisors engage with clients on easier to complete? So whether that’s paperwork or account, opening and onboarding, moving of assets, financial planning, these are all essential workflows related to how advisors interact with clients. To your earlier question, how are we using some of these kinds of AI technologies to help facilitate all this? It turns out that a lot of what I just said can be made easier, better faster, by leveraging AI so for example, if you’re an advisor, interacting with a client on Engage, Engage can actually help assist you in that meeting, whether it’s observing and analyzing what’s actually being discussed, and actually surfacing the right content, the right links the right information. So that the advisor can I can bring that as part of the conversation on paperwork, absolutely. The bank actually frequently has large amounts of data, in their middle and back office. If you can connect that data to the paperwork and map the fields, you can actually fill out the paperwork a lot more easily. The advisor does not sit there and type it all in.
Mike: There are obviously, we didn’t talk about conversational analysis and ChatGPT and all that stuff, but I don’t mean to downplay that, there will be lots of applications of language AI as it relates to, wealth and investing. That’s yet another area where, we’re seeing some interesting use cases, to essentially help the advisor run a better meeting. One of the things that’s interesting is some of the AI features are actually present during the meeting. So, I might be meeting with a client we’re talking about the fact that I just got married or had a kid or change jobs, and they I will actually be able to analyze the words being said and actually recommend different actions, different paperwork, different topics that the advisor might want us, engage with the client based on what the client is actually saying, but some of the AI actually is involved in helping the advisor before or after the beam.
Mike: One of the things that I think we’ve seen in the world is Salesforce has done an incredible job covering the market and getting lots of firms to spend lots of money getting onto Salesforce. They’re a great company, they’re great product, but one of the most common things that I see with firms who have adopted Salesforce or are in the process of adopting Salesforce, is that once they’ve spent all this money and they’ve implemented all this software, they step back after the first few months, and they realize that all of their hopes and dreams for what Salesforce can do for you, they’re not quite coming true.
One of the frequent failure points is that Salesforce has effectiveness in kind of collecting and tying all the data that you know the firm has about a client is completely reliant. On the data actually getting into the system. And when you think about real world usage of Salesforce, the best firms that adopt Salesforce optimally have some way of ensuring that advisors are actually putting all the data that they’re collecting into Salesforce. Unfortunately, the way the real world the way the real world works, advisors aren’t actually doing that very much. It’s a lot of effort, a lot of time a lot of work after a meeting is over. When you’re probably trying to run to your next meeting, or do your next thing, most people don’t take the half hour it takes to write it all up and type it all in.
Mike: One example of like where AI can be used after the meeting is through automated meeting summaries. If you run a meeting on Engage, we’re able to kind of analyze transcribe, record the meeting, we can actually synthesize the most important things that were happening in the meeting, we can prepare automated meeting notes and summaries for the advisor so they don’t actually have to write it all up in advance. And then to the extent any structured information was shared during the meeting, so you know, let’s say you are the advisor I was the client, you were asking me, oh, what’s happened in your life and how’s your family and this, that and the other.
I’m sharing all kinds of information that’s actually structured information, right. But again, that information does not end up in a structured way in the databases of most financial institutions. So what AI can do is actually automatically extract all of the structured information from the conversation and store it in a structured way so that all the other systems that might be able to benefit from leveraging, knowledge of my age or my wife’s age from how many kids I have, or how much money I have, or whatever is that piece of data, all of that data can kind of automatically kind of properly stored in the systems.
Craig: I can interrupt. You started out right when people think of SigFig as a wealth platform, right. As you mentioned, you started as a b2b robo, the built on traditional advisor technology all makes perfect sense. How did you get to this point, this is a very different capability that people wouldn’t normally think of SigFig as providing. So the product is called Engage, couple questions one, has it been launched? It was it in production, if it is elicited a lot of features and functionality, what features and functionality are available now? Which ones are on the roadmap?
Mike: Yes, it’s launched. How do we get here? I think we got here, partly by working closely with our partners, understanding what are they their pain points, where are they trying to go? And one of the patterns that we saw emerge, is there are a lot of firms that have very, very large audiences of mass affluent clients that they view as growth opportunities. Historically the wealth management firms have tended to focus on kind of high net worth clients. But for the banks and the insurance companies and the wealth management firms that we work with. There’s oftentimes a very large pool of mass affluent clients a step below kind of high net worth, and those clients represent high volume and in aggregate, represent large, AUM.
Mike: But in order to serve that client, well, you need a scalable, efficient model and so us being a technology company, it became quite clear that part of what we could do is build technology to create a better client experience. But also a more scalable business model in serving mass affluent clients.
Mike: When you look at the way firms want a certain mass affluent clients, one of the trends that you tend to see is that firms instead of having like the traditional field based advisor model, they’re moving towards more of like a centralized team based model context center base model. So you know, Vanguard was an early pioneer in this, they have this personal advisor services all of those advisors are like in a contact center centralized, they’re not out in bank branches or anything like that.
And again, one of the things that you see is that the type of advisor that they are trying to hire for that is usually a little bit younger, a little bit less experienced, and also a little bit more open to management guidance. As a traditional field data advisor, they’re usually a little bit older, the more experienced and more sophisticated and they’re sometimes resistant to kind of firm based, guidance on how they should be doing their job. And so all of that when you put it all together, with Engage, we realize that okay, we can build some technology that can make the interactions that advisors have with clients, much more effective, much more efficient. And we can leverage AI to create much more scalability in the way that advisors are actually interacting with those clients. And then for the home office, one of the things that you want is you want more consistency.
Mike: You think about the risk and compliance implications, the desire to have kind of a high quality service model. You don’t want to have like 1000 advisors where every advisor does something different. So all of that kind of walks back towards the idea of leveraging more modern technology to kind of run your business. That’s how we started, we started in robo advice. Then we expanded to kind of advisor technology, then we expanded to kind of virtual collaboration. And then it turns out that there are AI applications and all of those fields, we’ve kind of grown and expanded, over time just by following what are the biggest problems and pain points that we can solve for our partners?
Craig: The automated recording of client meetings, generating summaries extracting structured data, can I use that now? Is that available?
Mike: Yeah. So our teams have been building that we’re using that internally, and by the time this podcast goes live, it’ll probably be live in the market.
Craig: That’ll be incredible. If I can as an advisor, I want that as a consultant. I mean, I use AI tools to record my client meetings, but I’ve got to go back and do a quick summary. But it’s not going to send data to my CRM and create tasks and kick off workflows and things to get done. And also the summary needs tweaking the summaries are never good enough to send to the client. We always send a transcript on our projects of our calls with clients, reviewing deliverables a strategy and we send them an automated summary. We always have to tweak it. You’re talking to some pretty impressive capabilities, surfacing the right links and supporting data during the meeting. That means the company has to have their own private LLM that’s got all the source of knowledge in that company. So there’s a lot of other capabilities that need to be in place for this tool to work.
Early Adopters vs Pragmatists
Mike: One of the things actually that’s been nice about being a vertical software company is that we specialize and focus on financial services. You’ll see AI companies out there that are building kind of generalized AI across industries across domains. And that’s exciting. I’m sure that there’ll be a lot of fundamental AI that actually does work well across industries. But the nice part about being so focused on the Financial Services vertical, the wealth management vertical and working with financial advisors, specifically, is that we can decrease the complexity of the technology and increase the quality of the technology as it’s applied. Specifically to the problems that our industry faces. I think that that kind of accelerates the speed at which we can build useful things that actually work in the real world.
It’s funny at the very beginning of the podcast, you were saying how most people want to be on the cutting edge and it’s interesting, I think our experience with financial services is actually that some people want to be on the cutting edge, and they’re great as early adopters for our newest products, but many, many, many firms want things to be fully baked. They want to make sure it’s reliable, it’s secure, it’s stable. It’s private, all the regulatory requirements that our partners actually have actually caused them to oftentimes not want to be early adopters. And so having an environment where we can kind of beta test our capabilities and having our technologies being applied to very narrow spaces allows us to get to more mature products more quickly.
Craig: Absolutely, we work with a large broker dealers, lots of large insurance companies, and they’re clearly in a wait and see approach with some technology. They’re not going to just jump right in to feet in on AI as an example. We were talking to one of the top 10 insurance companies just the other day and their question was, how do we dip our toe in the water here? What’s the low hanging fruit that we can get some benefits from AI tools that won’t trigger regulatory or compliance issues? So things like meeting automation, you just automate as you’re taking existing information, automating it, summarizing it or expanding on it for advisors. That’s everything stays in house. Or it’s always run through compliance as all email is captured, is relatively safe, our call centers, working with your call center to give you’re building an LLM for all of the technology Technical Information Level One Call Center person might need and to be able to surface in for them for things like marketing, where you’re automating, generating social media content that still has to go to compliance, but it’s saving 80% of the time of doing the initial work.
Craig: Those are sort of low hanging fruit, where we’re seeing firms looking at at AI solutions and there’s more complicated ones, like you mentioned real time assistance. I saw a proof of concept from a custodian, wherever the chatbot where the advisor could say, my client lost their debit card, what do I do and it will come up saying, here’s the form you got to fill out it will populate the form with the clients information or or right away well, that’s that’s the hardest thing. You working with any large system whether you’re working at a large broker dealer, a wirehouse, custodian, just trying to find the myriad of options and things you need to do, as you mentioned 529, do a rollover or do someone passes away God forbid, whatsoever until the paperwork I need to do and having a system that can bring all that forward pre populate it move it all in the workflow for you would be incredibly beneficial for advisors understand.
Mike: Yeah, I agree. The topic you raised earlier which is like hey, if you if you’re interested in how you want to get started where like where are good places to start? I often refer to two concepts that I think are relevant in the answering of that question, where’s the low hanging fruit? One is, at the beginning, I may have mentioned when we think about AI, we think about it in three workstreams. One is technology that actually helps augment the ability of a human in a lot of circles, people are calling that copilot.
Mike: So, for example, there’s copilot for software engineers, where you’ve got aI helping you kind of validate that the code you wrote is optimal, or whatever it is, and so in our space, we’re big believers that humans will continue to be playing an important role in the delivery of financial services. And so using AI to augment the capability of the human is going to be very, very helpful. Unlike another second area, which is technology that actually helps automate away the need for a human right and so that obviously, also gets a lot of media attention is like, Oh, we’re all the jobs going to go the AI is gonna kind of put us all out at work. I think that’s going to happen to some extent in financial services. There’s a lot of stuff that’s done manually, and that shouldn’t be. So there’s gonna be a ton of efficiency gains by firms that adopt AI to replace kind of manual human processes.
Mike: Then the third angle is around customer experience. How can AI be used to create a better, more compelling customer experience that’s either better, faster, cheaper, some combination of all that. And so in those three buckets, I think from a early adopter regulatory risk perspective, the first bucket, technology that helps augment the capability of the human actually is probably the lowest risk area, because you always have the human being there, observing and analyzing for quality and being the belt and suspenders right on like an area if you’re focused on an area that was complete automation, eliminate the human, turn it all over to the machine, then you run some risks, what if the machine is wrong or what if it does something unexpected? So, that’s one framework, what is the domain space with the problem you’re trying to solve?
Mike: Then another area, for those, your audience who’s a little bit more interested in the underlying technology and techniques, one of the things that comes up a lot is that AI is built off models. And in the world of models, if you were to split the world of models into two, one way to split them, is this idea of statistical models versus computational models. And it turns out that one of these two, I think, has lower risk versus higher risk. Statistical models are ones that are built off analyzing large datasets. And what they try to do is they try to predict based on historical patterns, what is more or less likely. For example, marketing optimization, there’s no right or wrong answer there are more likely to be effective, less likely effective marketing strategies, fraud analysis.
Mike: This is another area that AI has been used for many, many years, one of my first jobs was actually building fraud prediction models and fraud detection models for Amazon. That’s another area where you’re using big datasets and statistics to try to predict the likelihood of fraud versus computational models. These are models that are built around the idea that there is a right and a wrong, if you built a model to to analyze what what two plus two is, there’s a right model and the wrong model. I think this is an area where like, a lot of people believe that in the long long term, there will be fully optimized you know, self driving finance algorithms that help clients optimize how their money is managed, that don’t involve much human intervention. That’s kind of a more of a computational model. And if you think about it, the regulatory risk implications of computational models being wrong are incredibly high.
Mike: If you’re a financial institution and you adopted AI to give your clients financial advice, and it turns out the AI was giving bad advice, that would be incredibly risky bad reputational damage, your regulators would be all over you. The idea of kind of low hanging fruit work and I start, I generally tend to point people to more statistical use cases where, there’s less chance of being right versus wrong, and more tolerance for error is not even the right word. Because there is no right or wrong, it’s more about creating kind of the right predictions that help optimize outcomes. Those are some of the areas that we oftentimes, when we think about the longer term arc of AI and where do we invest in the near term versus longer term investments, we’re placing more of our investments in areas that are likely to be commercially valuable sooner.
Craig: That makes perfect sense. Why, you’ve got to have some things that are coming out soon. You can’t put everything in the long term budget. You’ve got to have some of your tech budget towards functionality that you can sell now. So, to summarize, we’ve got some new technology coming. We got some stuff in the pipeline. It’s going to be coming out very soon. We’re kind of out of time, I wanted to kind of dig a little bit farther, but time just flies here. Can you tell the audience where they can find more information about SigFig?
Mike: Just visit our website, SigFig.com. Obviously, if you’re an institution, we’ve got lots of great team members who would love to talk to you and share with you demos and examples of the way we’re working with many different firms. If you’re an advisor interested in this, obviously we don’t work with individual advisors directly, but we do work with lots of advisors through our partners. That’s part of how we work is we work with kind of institutions. If anyone wants to contact us, you got contact info and happy to share it with you so that your listeners can reach out to us.
Craig: Mike, thanks so much for being here.
Mike: My pleasure and great to be on the show and look forward to keeping in touch over time.